Joho the Blog
An Entry from the Archives

« [massnetcomms] Net neutrality panel || Back to Blog | One Web Day at the Berkman Center »

September 08, 2006

DOEP - The Daily Open-Ended Puzzle (one time only)

Say a city decides to put its public bus stops on the sides of the intersection immediately after the traffic lights rather than immediately before the lights. Will this:

a. Decrease the overall transit time from one side of the city to another

b. Increase the overall transit time

c. Have no consistent effect on overall transit time

d. Go away you boring git

(Note: I'm not claiming to have an answer to this.) [Tags: puzzle quiz doep]

Posted by D. Weinberger at September 8, 2006 11:28 AM


Comments

That depends on whose transit time you have in mind - that of motorists, bus riders, or the bus driver. The bus driver's transit time will be less, because although he still stops at each stop, he is less likely to be caugth by a red light immediately after a stop. The bus riders' time, on average, will increase. This is because some percentage of riders always find themselves on the wrong side of the street as they race to catch the bus, but now they will need to cross against traffic to get to the stopped bus, (which reached the stop by going through a green light), and that necessity will cause more of those unfortunate riders to miss the bus. It's a wash for motorists - the ones talking on their cell phones will still get stuck behind the bus while the alert ones will find a way round. I prefer walking myself.

Posted by: DavidT | September 8, 2006 01:36 PM


No wonder I'm always missing buses! There may be five stops within a block from me, but four of them are across the street, immediately after the intersection. Thanks, DavidT, for making me realize what was happening.

Posted by: Simone | September 8, 2006 01:49 PM


e - where are my fraking fly cars, weren't we promised flying cars in the 1960s by now. If I had my flying car I could zip around the bus!

Posted by: Thomas | September 8, 2006 02:00 PM


Given random lights, and random bus times, doesn't the following hold:

1. Any bus will be stopped at the red lights randomly, based on whatever green/red cycle is being used.
2. So a bus approaching a red light will be stopped there some proportion of the time. Whether the people get on/off before or after the light makes no difference. The likelihood of being stopped at the light is the same in either case, since the probability of getting stopped is the same in either case.
3. We can extrapolate, therefore, that it makes no difference.

A real improvement would be to allow dropping people off at the corner before the light if and only if the light is red as the bus approaches, but allow boarding only after the light. But the benefit might be marginal. A single rider might save a minute of time by getting out at a red light, but if the ride is 25 minutes long, that's not much of a savings. In a city with short commutes and long red lights, it might pay off.

Posted by: Stowe Boyd | September 8, 2006 02:20 PM


I'm with Stowe: I don't think it makes a difference. FWIW, here in Portland (OR), the drivers will let you off at the red light before your stop (and most stops are in fact after the light), if it's safe to do so (meaning, they're next to the curb not one lane out).

Posted by: Bill Hooker | September 8, 2006 06:52 PM


In general, it seems that pre-light stops are a bad idea if there's a lot of traffic (meaning that the bus can only pull up to the stop when traffic starts moving on a green light) and the average time to load passengers is about the length of a green cycle. They're also a bad idea because of the all-too-familiar "newly-arrived passenger banging on the doors just as the light turns green" problem.

My analysis depends on the assumption that conditions are chaotic between lights (high variation in mean time to reach the next stop, etc), and that a bus stopped at a light behind other cars takes negligible time to pull up to a pre-light stop. I also assume that we're talking about a bus-centric world here, where we care only about the transit time of the bus passengers.

Given those assumptions (which seem reasonable), and the conditions under which pre-light stops are a bad idea (which also sound plausible), I tend to believe that the answer is A.

Note that, if all other conditions are held constant but the average time to load passengers is instead about the length of a full green+red cycle, then the answer might be B. But we still have the late-arriving-passenger problem, which mitigates against B.

In a world with no other traffic and no late-arriving passengers, I think the answer is C.

A is my final answer.

Posted by: Matt Norwood | September 8, 2006 07:53 PM


In reference to Stowe's answer above: I think it abstracts away the phenomenon of traffic "spurts" and the fact that a bus stopped at a light may be so far from its stop that it has to wait for green, then pull up to the stop and wait for passengers to board. In this kind of system, buses are much more likely to arrive at lights that are green than lights that are red.

Of course, we may just have very different trafic conditions in mind.

Posted by: Matt Norwood | September 8, 2006 08:00 PM


Post a comment

Guidelines for Commenting

Basically, you can say what you want. (Click here for the fine print.)

If you haven't left a comment here before, your comment may be put into a queue for me to approve. Sorry for the delay. Blame the damn spammers.